This addresses #13441.
This preserves the similar check at the _point of use_ of invalid messages in `DescriptorBuilder` (and there's an existing test that verifies that check still works).
But it adds another check in the parser, to catch this error at the _point of definition_ of an invalid message. And the corresponding test is updated: we no longer need a usage of the message to catch the error, and the reported position is the definition of the option, not the usage site of the message.
The way this works feels kinda gross, but I wasn't sure of a better way to do it. The only place we know for certain that it was an explicit option (vs. auto-added by the parser when synthesizing a map entry message) is when after processing the message body, we can look at the uninterpreted options. So that's what this does. If you have ideas on better/cleaner approaches, I'd be happy to revise.
Closes#13479
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/13479 from jhump:jh/map-option-not-allowed 4d95e9b839
PiperOrigin-RevId: 555577734
This is in preparation for moving upb into the protobuf repo. Upb requires C++
17. Protobuf supports C++ 14, and our non-upb CI tests will keep using C++ 14
to ensure that it still works. Our Bazel CI runs use the bazelrc files in the
ci/ directory, so these runs should be unaffected by this change.
PiperOrigin-RevId: 555223181
Only emit has_field() if the field support presence. Only emit field_opt() getter if the field is both optional and supports presence.
PiperOrigin-RevId: 555133374
With some optimizer flags, GPBCodedOutputStream results in a lot of
code bloat, for now just prevent the new helper related to
NSOutputStream from being inlined to keep that in check.
PiperOrigin-RevId: 554525631
When build a FileDescriptorProto into pool, FieldDescriptorProto.type may not set if type_name is set. This only add tests. UPB python Fix has been submit in other change
PiperOrigin-RevId: 554516745
Add some missing endian conversions so that the conformance tests can be run on big endian platforms.
The message length value created by the conformance test runner is little endian according to the comments in the file but actually was sent in the native endianness of the host. I was able to run the java, python, ruby, php and csharp test executables and they all expect little endian length values so those tests would hang on big endian machines. Only the cpp test executable was using native endian so it has been changed to expect little endian too.
Also change the fixed32 and fixed64 functions in binary_json_conformance_test_suite.cc to send the data as little endian which fixes some failures in the python conformance tests on big endian platforms.
Closes#13443
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/13443 from linux-on-ibm-z:conformance-runner-little-endian-fix 4ef7948997
PiperOrigin-RevId: 553958649
Change the pack/unpack format codes 'f' -> 'g' and 'd' -> 'e' which ensure little endian format is used on the wire regardless of machine endianness. This allows the php composer and conformance tests to pass on big endian platforms and should not change any behavior on little endian platforms.
According to the PHP documentation, 'g' and 'e' were added in PHP versions 7.0.15 and 7.1.1.
Closes#13444
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/protobuf/pull/13444 from linux-on-ibm-z:php-endian-fix 7840e09755
PiperOrigin-RevId: 553954880
This CL adds function bodies for: {as_view, into_view, as_mut, into_mut, set_on} [1].
Our prior cl/552609955 didn't have `RawMessage` inside $Msg$Mut, so that's also been rectified in this CL.
[1] Everything in this set is a part of the Proxied trait, except set_on, which belongs to the SettableValue trait.
PiperOrigin-RevId: 553935803
do them only once at the beginning instead of for each element.
In the past we hoisted the `rep` checks. This change continues that pattern.
PiperOrigin-RevId: 553901961
When build a FileDescriptorProto into pool, FieldDescriptorProto.type may not set if type_name is set. Runtime Change. To make copybara happy, tests will be added in a separate change
PiperOrigin-RevId: 553881407
The file structure was broken only when invoking from protobuf repo (`bazel build @upb//python/dist:source_tarball`) since filegroup files have a path like `protobuf/external/upb/python/dist/setup.py` rather than `protobuf/setup.py`
This was broken by 79af13abde
PiperOrigin-RevId: 553821217
Clang and GCC differ on how they detect Address Sanitizer. Support both.
Closes#1424
COPYBARA_INTEGRATE_REVIEW=https://github.com/protocolbuffers/upb/pull/1424 from protocolbuffers:asan-clang 491a5ee4cfd24c8eb281f894de0cf4384525c46a
PiperOrigin-RevId: 553805994
When build a FileDescriptorProto into pool, FieldDescriptorProto.type may not set if type_name is set. Runtime Change. To make copybara happy, tests will be added in a separate change
PiperOrigin-RevId: 553598520
A write succeed but only write out a subset of the requested amount
of data, handle that via a helper that will loop and advance as need
to completely write out the data.
PiperOrigin-RevId: 553516360
Joshua Humphries
Sandy Zhang
Protobuf Team Bot
Protobuf Team Bot
ericsalo
Mike Kruskal
Adam Cozzette
Eric Salo
Mike Kruskal
Jie Luo
Thomas Van Lenten
Matt Kulukundis
Jonathan Albrecht
Euclid
Hong Shin
Jason Lunn